Why We Chose Algorithms (OR-Tools) Over AI?
One of the most frequent questions we get is: "In the era of LLMs, why didn't you just use AI to generate the schedules?"
While AI is a powerful "Copilot," using it as an "Autopilot" for school scheduling is like trying to solve a Rubik's cube with a paintbrush.
Deterministic vs. Probabilistic Results
AI models are probabilistic. They predict the next most likely token. When you have 120 classes and 500 teachers, you don't need a "likely" schedule; you need a guaranteed one. AI can suffer from hallucinations. It might "forget" that a teacher can't be in two places at once or hallucinate a 25th hour in a day.
Computational Efficiency (Single-Core Friendly)
Modern AI requires massive GPU clusters and gigabytes of VRAM to function effectively. We believe school tech should be accessible, without any technological and financial gaps.
By using Python-based discrete logic, the mathematical solver is incredibly lightweight. This algorithmic approach can calculate complex schedules on a single-core computer or a standard laptop. You don't need an entire datacentre to power your timetables up, wasting tons of computation power.
AI is the Copilot, Not the Autopilot
During development, we actually tried using AI to write the scheduling logic. The result? The code was unusable. We spent more time fixing the AI's "logic leaps" than we did writing the actual system.
While AI is now a trend (I mean, most project are now incoporates with AI system), it doesn't means that it is a must, sometimes solving problems in the so-called "traditional" way can be much more efficient than asking AI to do everything for you.
Remember this slogan
AI is the copilot, not the autopilot!
Want to see the math in action?
Check out our solver/ directory to see how we translate human needs into discrete logic predicates.
Log in or sign up for Devpost to join the conversation.